People have attempted to simulate human voices in machines for quite some time. In the late 1700s, Hungarian scientist Wolfgang

admin2019-07-05  46

问题     People have attempted to simulate human voices in machines for quite some time. In the late 1700s, Hungarian scientist Wolfgang yon Kempelen used whistles, resonance chambers, and other objects to create voice sounds for his "Speaking Machine. " With the rise of computers, scientists were able to create artificial voices by storing and digitally processing real voice sounds.  By the 1970s, these computer-generated voices were widely used. Although these early voices were understandable, they had a distinct mechanical or " robotic" sound to them.
    Coming up with more natural-sounding voices has been a challenge due to the complex nature of language. (86) To produce a natural-sounding voice, scientists need to simulate not only the individual sounds of a language, but also the volume, pitch, rhythm, and tones that help to express meaning. Natural-sounding computer voices are now used to provide information to people in a variety of applications, from automated phone lines to language learning programs. Many systems are also able to listen to users’ questions and statements and respond to them.
    One good example of this comes from the BMW car company. BMW has programmed cars to both listen and speak to the driver. (87) The car can give directions, provide warnings and information about traffic and safety conditions, and even control certain functions, such as raising or lowering the windows or playing music. When BMW first released cars with a computer-generated voice, many drivers had a negative reaction to it because drivers perceived the voice as female.  German drivers were uncomfortable with a "female" voice giving directions. As a result, BMW recalled the cars and gave them "male" voices.
    The reaction to the female voice was emotional. It had no reasoning behind it. Common sense says that the "gender" of the voice in the car shouldn’t matter because drivers know that they are listening to a computer, not a person. (88) But even without a visual representation of the voice, such as a male or female face, people reacted with the same stereotypes they would apply to a person. Studies have also shown that people react similarly to other aspects of’ computer voices, such as the "personality" of the voice or its level of "politeness. "
    Researchers now know that voice interfaces are actually social interfaces; that is, people will react to a computer using the same rules and expectations that they would apply to people. (89) As a result, designers of these systems consider not only the sounds of the voices they use, but also their psychological effect on users.

选项

答案为了发出自然的声音,科学家们不仅需要模仿语言中每一个单独的音,也需要模仿帮助表达意义的音量、声调、节奏和语气。

解析 (not only…but also不仅…而且。natural-sounding听起来自然的。simulate模仿。)
转载请注明原文地址:https://jikaoti.com/ti/QZhMFFFM
0

最新回复(0)